Oxidation Suppression During Cu

3.4.1 The Impact of Initially Dissolved Oxygen

In an effort to reduce the synthesis complexity, we aimed at defining synthe-sis conditions for which Cu⁰ NCs are formed under ambient. In principle, this would require us to strictly avoid exposure to oxygen at all stages of the synthesis. As a first oxygen source, we considered all oxygen initially dis-solved in the OlNH2-dodecane solvent mixture at the start of the reaction.

In chemical laboratories, it is common practice to protect oxygen-sensitive compounds during a reaction by degassing the initial reaction mixture under vacuum. However, degassing is a time-consuming process that is impracti-cal for large volumes. Moreover, if a reaction under ambient atmosphere is envisaged, initial degassing seems futile since oxygen can dissolve back into the solvent prior to the reaction.

To evaluate the impact of dissolved oxygen on the formation of Cu NCs, we estimated the concentration of dissolved oxygen in the initial OlNH2 -dodecane solvent mixture. According to published data,¹⁸the solubility of oxygen in n-dodecane amounts to 54.9 mg/L at 35^◦C and atmospheric pres-sure. Using that value, we calculated that for the 2 g/L synthesis described in the previous section, the molar ratio between Cu formate and dissolved O2 amounts to∼18. To interpret the consequence of such an equivalence, we considered the formation of Cu₂O from Cu and O₂:

4Cu(s)+ O2(sol)2Cu2O(s) (3.1) According to Equation 3.1, 4 moles of Cu are oxidized by 1 mole of O₂. Therefore, a sizable fraction of about 20% of the reduced copper could be re-oxidized by the oxygen initially dissolved in the reaction mixture of that 2 g/L synthesis. This suggests that further increasing the Cu(HCO2)2 con-centration in the reaction mixture could be an efficient strategy to suppress oxidation. Indeed, such reaction mixtures would have an increased initial Cu/O2 ratio and, consequently, a smaller fraction of re-oxidized Cu as a final product. For example, an increase of the Cu content to 40 g/L could decrease the expected fraction of re-oxidized copper to less than 1% of the total amount of Cu formed. Using such an approach, the initial oxidation

Intensity (a.u.)

Figure 3.3: XRD patterns of Cu NCs synthesized with different con-centrations of Cu without protective atmosphere. The molar ratio of OlNH2:Cu(HCO2)2has been kept at 2:1 and the heating rate at20^◦C/min for all syntheses. The NCs were purified and dried in ambient conditions prior to measurement.

Table 3.1: Copper ratios extracted from Rietveld refinement analysis of the XRD data from Figure 3.3.

[Cu] Cu (%) Cu2O(%)

of this small amount of Cu may lead to an oxygen-free reaction mixture;

a self-cleaning effect that creates the inert environment necessary for the formation of metallic Cu NCs.

We tested this possible self-cleaning of the reaction mixture by running the Cu(HCO2)2 decomposition under ambient conditions in reaction mixtures with a different solid loading. Figure 3.3 presents the XRD patterns of the thus obtained NCs, for Cu loadings increasing from 2 to 50 g/L. As can be seen, the initial concentration of the copper precursor has a large impact on the oxidation state of the resulting NCs. At low precursor concentration, the NCs mainly consist of Cu2O. In synthesis conditions expected to pro-duce 2 g/L of Cu, the XRD pattern obtained is similar the data presented in 3.1c since the synthesis conditions are equivalent. As the concentration of the initial precursor is increased, the crystallographic phases gradually shift from Cu2O to Cu⁰. Table 3.1 summarizes the results of a Rietveld

refinement of these diffractograms, showing that the metallic Cu content increases from 8.8% to 97.3% by increasing the concentration from 2 g/L to 40 g/L. Note that the oxidized fraction is about 2 % for the 40 g/L synthe-sis, close to the estimated value based on the room temperature solubility of oxygen in dodecane. As an additional benefit, a larger concentration of precursor results in a higher amount of material produced per volume of reaction; synthesis conditions that reduce the overall cost of producing Cu NCs and limit the generated waste.

3.4.2 Oxygen intake during synthesis

Apart from the initially dissolved oxygen, the influx of oxygen from the ambient surroundings can be a second oxygen source in an open reaction vessel, which we succinctly describe as a transition from gaseous to dissolved oxygen:

O_2(g)O_2(sol) (3.2)

This process will promote redissolution of oxygen as already dissolved oxy-gen is consumed by the oxidation of Cu⁰ as described by Eq 3.1. Hence, to prevent further formation of copper oxide, the influx of oxygen in the reac-tion mixture should be restricted during the decomposireac-tion of the precursor and the nucleation and growth of Cu NCs. However, the flask cannot simply be sealed to block any gas inflow since the decomposition of Cu(HCO2)2

releases water, carbon dioxide and hydrogen:⁹

Cu(HCO₂)₂·4H₂O_(sol)→Cu_(s)+ H_2(g)+ 2CO_2(g)+ 4H₂O_(g) (3.3) Clearly, such gas evolution would result in a problematic pressure increase in a closed flask.

To evaluate the impact of redissolved oxygen on the reaction, we started from the idea that gasses released from the reaction mixture as Cu(HCO2)2

decomposes can suppress oxygen redissolution in the reaction mixture. In-deed, considering that the reaction occurs in an open flask exposed to ambi-ent atmosphere, the total pressure in the headspace must remain constant.

When the decomposition of Cu(HCO₂)₂ starts, evaporation of water and the generation of hydrogen and carbon dioxide from the solution increase the partial pressure of such gases above the liquid and decrease the oxygen partial pressure to keep the total pressure constant. According to Henry’s law, a lower partial pressure of oxygen implies that less oxygen will dissolve in the reaction mixture, thus protecting the Cu NCs formed from oxidation.

Intensity (a.u.)

Figure 3.4: (a) XRD patterns of Cu NCs synthesized at different heating rates without protective atmosphere, to a final temperature of140^◦C. (b) XRD patterns of Cu NCs synthesized without protective atmosphere, at 10^◦C/min to different final temperatures and held at these temperature for 10 min. The concentration of copper was set at 20 g/L and the molar ratio ofOlNH2:Cu(HCO2)2 2:1 for all syntheses. The NCs were purified and dried in ambient conditions prior to characterization.

Importantly, the formation of such an oxygen-poor atmosphere in the reac-tion vessel will strongly depend on the reacreac-tion rate, where slow reacreac-tions will have hardly any impact, whereas fast reactions may strongly influence oxygen redissolution.

We evaluated the impact of the reaction rate on the formation of copper and copper oxide during Cu(HCO2)2decomposition by either changing the rate at which the a final reaction temperature of 140^◦C is reached, or by changing that reaction temperature. Figure 3.4 illustrates the x-ray diffrac-togram recorded on the reaction product obtained for these different syn-theses, which give evidence of an increased copper oxide content for slower heating rates or lower reaction temperatures, i.e., conditions in which the overall reaction rates is lowest. This qualitative picture is confirmed by a Rietveld analysis, the results of which are summarized in Tables 3.2 and 3.3. One sees that at the lowest heating rate used, a mixed copper/copper oxide product is obtained whereas a heating rate at 20 ^◦C/min leads to a

Table 3.2: Copper ratios extracted from Rietveld refinement analysis of the XRD data from Figure 3.4a.

Heating rate (^◦C/min) Cu (%) Cu2O(%) 2.5 43.5±0.7 56.5±0.7 5 77.6±1.2 22.4±1.2 10 82.5±0.8 17.5±0.8 20 92.4±1.6 7.6±1.6

Table 3.3: Copper ratios extracted from Rietveld refinement analysis of the XRD data from Figure 3.4b.

Final temperature (^◦C) Cu (%) Cu₂O(%)

120 8.5±0.7 91.5±0.7

140 82.5±0.8 17.5±0.8

160 90.1±4.4 9.9±4.4

NC product composed for ∼92% of Cu⁰. Note that the latter result is in close agreement with the data listed in Table 3.1 for the same precursor concentration. Similarly, Table 3.3) indicates that changing the reaction temperature from 120^◦C to 160^◦C induces a shift of the end product com-position from 90% copper oxide to 90% copper. Hence, we conclude that a lower reaction rate strongly promotes the formation of copper oxide, proba-bly by enabling redissolution of ambient atmosphere in the reaction mixture.

On the other hand, high reaction rates can effectively suppress the redisso-lution of oxygen, which lead to an oxygen-poor reaction mixture in which Cu NCs can be formed, even under ambient conditions.

3.5 Size-Tuning During Cu Nanocrystal